CadAI-B Dx

K260086 · BeamWorks, Inc. · QDQ · Aug 12, 2026 · Radiology

Device Facts

• Record ID: K260086
• Device Name: CadAI-B Dx
• Applicant: BeamWorks, Inc.
• Product Code: QDQ · Radiology
• Decision Date: Aug 12, 2026
• Decision: SESE
• Submission Type: Traditional
• Regulation: 21 CFR 892.2090
• Device Class: Class 2
• Attributes: AI/ML, Software as a Medical Device, Real-World Evidence

Real-World Evidence

Submission	Device	Sponsor	RWD Sources	RWE Use Summary	Key Tags
K260086 · Aug 12, 2026	CadAI-B Dx	BeamWorks, Inc.	Retrospective clinical breast ultrasound image databases; Radiology and pathology reports	Retrospective clinical cases were used to conduct a Multiple Reader Multiple Case (MRMC) study and standalone performance testing to validate the device's diagnostic accuracy and clinical effectiveness in assisting physicians with breast lesion localization and malignancy risk analysis.	Retrospective clinical data; Clinical validation; Breast ultrasound; Diagnostic accuracy

Clinical Evidence

Study Design	Population	Comparator	Key Endpoints
MRMC Reader Study; Multiple Reader Multiple Case (MRMC) study; Study Period: 2011–2023 (image acquisition years)	Patients with soft tissue breast lesions referred for diagnostic ultrasound; Sample Size: 797 cases (447 benign, 350 malignant); Number of Sites: Multiple clinical sites	Unaided reading condition	AULROC, Sensitivity, Specificity, PPV, NPV
Standalone Performance Evaluation; Retrospective standalone performance testing	Patients with soft tissue breast lesions; Sample Size: 1,285 cases (603 benign, 682 malignant); Number of Sites: Multiple (U.S., South Korea, Kazakhstan)	Not applicable for this study	AULROC, Sensitivity, Specificity, PPV, NPV

AI Performance

Output	Algorithm	Acceptance	Observed	Dev DS	Dev Readers	Test DS	Test Readers
Breast lesion malignancy risk analysis	AI/ML algorithms	AULROC > 0.7626	AULROC 0.907	—	—	Standalone performance evaluation: 1,285 cases	—
Breast lesion malignancy risk analysis	AI/ML algorithms	—	AULROC 0.873 (Aided) vs 0.766 (Unaided)	—	—	MRMC reader study: 797 cases	16 (readers)

Indications for Use

CadAI-B Dx is a software application indicated to assist interpreting physicians in analyzing the breast ultrasound images of patients with soft tissue breast lesions suspicious for breast cancer who are being referred for further diagnostic ultrasound examination. Output of the device includes lesion indicators placed on breast ultrasound images assisting physicians to identify suspicious soft tissue breast lesions from given B-Mode ultrasound images and the risk analysis of malignancy. The risk analysis of malignancy indicates the malignancy score (CadAI-Score) and corresponding BI-RADS categories. In addition, CadAI-B Dx analyzes the size of the lesions and BI-RADS lexicon descriptors: shape, orientation, margin, echo pattern, and posterior features. CadAI-B Dx may also be used as an image viewer of digital ultrasound images. The software includes tools that allow users to adjust, measure and document images, and output into a structured report. Patient management decisions should not be made solely on the basis of analysis by CadAI-B Dx. Limitations: CadAI-B Dx is not to be used on sites of post-surgical excision, or images with Doppler, elastography, or other overlays present in them.

Device Story

CadAI-B Dx is a SaMD that provides CADe and CADx for breast ultrasound. It takes static B-Mode ultrasound images (DICOM/RGB) as input. Using AI/ML algorithms, it identifies and highlights suspicious soft tissue lesions, calculates a malignancy score (0-100), suggests BI-RADS categories, and extracts BI-RADS lexicon descriptors (shape, orientation, margin, echo pattern, posterior features). It also provides automated lesion size measurements. Used in clinical settings by interpreting physicians, the device acts as an image viewer with tools for adjustment, measurement, and structured reporting. Results are saved as DICOM Secondary Capture files to PACS. The AI output is gated via role-based access control. By providing objective lesion localization and malignancy risk analysis, the device assists physicians in diagnostic interpretation, potentially improving diagnostic accuracy, reducing false positives, and enhancing workflow efficiency. It is not intended for primary interpretation of mammography or use on mobile devices.

Clinical Evidence

Clinical validation used an MRMC study with 16 readers and 797 cases (447 benign, 350 malignant). Primary endpoint was AULROC. Reader-averaged AULROC improved from 0.766 (unaided) to 0.873 (aided) (p < 0.001). Sensitivity improved from 91.29% to 94.23% (p=0.015); specificity improved from 36.09% to 55.34% (p < 0.001). Standalone performance testing on 1,285 independent cases yielded an AULROC of 0.907 (95% CI: 0.890-0.924), sensitivity of 88.12%, and specificity of 77.11%.

Technological Characteristics

Software-only device operating on off-the-shelf hardware. Analyzes static B-Mode ultrasound images (DICOM/RGB). Employs AI/ML algorithms for lesion detection, classification, and quantification. Features include role-based access control for diagnostic outputs, DICOM Secondary Capture generation, and structured reporting. Complies with ISO 14971 for risk management.

Indications for Use

Indicated for patients with soft tissue breast lesions suspicious for breast cancer referred for diagnostic ultrasound. Contraindicated for use on sites of post-surgical excision or images containing Doppler, elastography, or other overlays.

Regulatory Classification

Identification

A radiological computer-assisted detection and diagnostic software is an image processing device intended to aid in the detection, localization, and characterization of fracture, lesions, or other disease-specific findings on acquired medical images (e.g., radiography, magnetic resonance, computed tomography). The device detects, identifies, and characterizes findings based on features or information extracted from images, and provides information about the presence, location, and characteristics of the findings to the user. The analysis is intended to inform the primary diagnostic and patient management decisions that are made by the clinical user. The device is not intended as a replacement for a complete clinician's review or their clinical judgment that takes into account other relevant information from the image or patient history.

Special Controls

A radiological computer assisted detection and diagnosis software must comply with the following special controls: Design verification and validation must include: 1. i. A detailed description of the image analysis algorithm, including but not limited to a description of the algorithm inputs and outputs, each major component or block, how the algorithm and output affects or relates to clinical practice or patient care, and any algorithm limitations. ii. A detailed description of pre-specified performance testing protocols and dataset(s) used to assess whether the device will provide improved assisted-read detection and diagnostic performance as intended in the indicated user population(s), and to characterize the standalone device performance for labeling. Performance testing includes standalone test(s), side-by-side comparison(s), and/or a reader study, as applicable. iii. Results from standalone performance testing used to characterize the independent performance of the device separate from aided user performance. The performance assessment must be based on appropriate diagnostic accuracy measures (e.g., receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Devices with localization output must include localization accuracy testing as a component of standalone testing. The test dataset must be representative of the typical patient population with enrichment made only to ensure that the test dataset contain a sufficient number of cases from important cohorts (e.g., subsets defined by clinically relevant confounders, effect modifiers, concomitant disease, and subsets defined by image acquisition characteristics) such that the performance estimates and confidence intervals of the device for these individual subsets can be characterized for the intended use population and imaging equipment. iv. Results from performance testing that demonstrate that the device provides improved assisted-read detection and/or diagnostic performance as intended in the indicated user population(s) when used in accordance with the instructions for use. The reader population must be comprised of the intended user population in terms of but not limited to clinical training, certification, and years of experience. The performance assessment must be based on appropriate diagnostic accuracy measures (e.g., receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Test datasets must meet the requirements described in 1(iii) above. v. Appropriate software documentation, including device hazard analysis, software requirements specification document, software design specification document, traceability analysis, system level test protocol, pass/fail criteria, testing results, and cybersecurity measures. 2. Labeling must include the following: i. A detailed description of the patient population for which the device is indicated for use. ii. A detailed description of the device instructions for use, including the intended reading protocol and how the user should interpret the device output. iii. A detailed description of the intended user, and any user training materials as programs that addresses appropriate reading protocols for the device to ensure that the end user is fully aware of how to interpret and apply the device output. iv. A detailed description of the device inputs and outputs. v. A detailed description of compatible imaging hardware and imaging protocols. vi. Warnings, precautions, and limitations must include situations in which the device may fail or may not operate at its expected performance level (e.g., poor image quality or for certain subpopulations), as applicable. vii. A detailed summary of the performance testing, including: test methods, dataset characteristics, results, and a summary of sub-analyses on case distributions stratified by relevant confounders, such as anatomical characteristics, patient demographics and medical history, user experience, and imaging equipment.

*Classification.* Class II (special controls). The special controls for this device are:(1) Design verification and validation must include: (i) A detailed description of the image analysis algorithm, including a description of the algorithm inputs and outputs, each major component or block, how the algorithm and output affects or relates to clinical practice or patient care, and any algorithm limitations. (ii) A detailed description of pre-specified performance testing protocols and dataset(s) used to assess whether the device will provide improved assisted-read detection and diagnostic performance as intended in the indicated user population(s), and to characterize the standalone device performance for labeling. Performance testing includes standalone test(s), side-by-side comparison(s), and/or a reader study, as applicable. (iii) Results from standalone performance testing used to characterize the independent performance of the device separate from aided user performance. The performance assessment must be based on appropriate diagnostic accuracy measures ( *e.g.,* receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Devices with localization output must include localization accuracy testing as a component of standalone testing. The test dataset must be representative of the typical patient population with enrichment made only to ensure that the test dataset contains a sufficient number of cases from important cohorts (*e.g.,* subsets defined by clinically relevant confounders, effect modifiers, concomitant disease, and subsets defined by image acquisition characteristics) such that the performance estimates and confidence intervals of the device for these individual subsets can be characterized for the intended use population and imaging equipment.(iv) Results from performance testing that demonstrate that the device provides improved assisted-read detection and/or diagnostic performance as intended in the indicated user population(s) when used in accordance with the instructions for use. The reader population must be comprised of the intended user population in terms of clinical training, certification, and years of experience. The performance assessment must be based on appropriate diagnostic accuracy measures ( *e.g.,* receiver operator characteristic plot, sensitivity, specificity, positive and negative predictive values, and diagnostic likelihood ratio). Test datasets must meet the requirements described in paragraph (b)(1)(iii) of this section.(v) Appropriate software documentation, including device hazard analysis, software requirements specification document, software design specification document, traceability analysis, system level test protocol, pass/fail criteria, testing results, and cybersecurity measures. (2) Labeling must include the following: (i) A detailed description of the patient population for which the device is indicated for use. (ii) A detailed description of the device instructions for use, including the intended reading protocol and how the user should interpret the device output. (iii) A detailed description of the intended user, and any user training materials or programs that address appropriate reading protocols for the device, to ensure that the end user is fully aware of how to interpret and apply the device output. (iv) A detailed description of the device inputs and outputs. (v) A detailed description of compatible imaging hardware and imaging protocols. (vi) Warnings, precautions, and limitations must include situations in which the device may fail or may not operate at its expected performance level ( *e.g.,* poor image quality or for certain subpopulations), as applicable.(vii) A detailed summary of the performance testing, including test methods, dataset characteristics, results, and a summary of sub-analyses on case distributions stratified by relevant confounders, such as anatomical characteristics, patient demographics and medical history, user experience, and imaging equipment.

Predicate Devices

• BU-CAD (K210670)

Submission Summary (Full Text)

{0}

**FDA** U.S. FOOD & DRUG
ADMINISTRATION

BeamWorks, Inc.
MinGi Seo
Regulatory Affairs Manager
B1, 107, Chilgokjungang-Daero 136-Gil, Buk-Gu
Daegu, 41404
Republic Of Korea

August 12, 2026

Re: K260086

Trade/Device Name: CadAI-B Dx

Regulation Number: 21 CFR 892.2090

Regulation Name: Radiological Computer-Assisted Detection And Diagnosis Software

Regulatory Class: Class II

Product Code: QDQ

Dated: July 13, 2026

Received: July 13, 2026

Dear MinGi Seo:

We have reviewed your section 510(k) premarket notification of intent to market the device referenced above and have determined the device is substantially equivalent (for the indications for use stated in the enclosure) to legally marketed predicate devices marketed in interstate commerce prior to May 28, 1976, the enactment date of the Medical Device Amendments, or to devices that have been reclassified in accordance with the provisions of the Federal Food, Drug, and Cosmetic Act (the Act) that do not require approval of a premarket approval application (PMA). You may, therefore, market the device, subject to the general controls provisions of the Act. Although this letter refers to your product as a device, please be aware that some cleared products may instead be combination products. The 510(k) Premarket Notification Database available at https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm identifies combination product submissions. The general controls provisions of the Act include requirements for annual registration, listing of devices, good manufacturing practice, labeling, and prohibitions against misbranding and adulteration. Please note: CDRH does not evaluate information related to contract liability warranties. We remind you, however, that device labeling must be truthful and not misleading.

If your device is classified (see above) into either class II (Special Controls) or class III (PMA), it may be subject to additional controls. Existing major regulations affecting your device can be found in the Code of Federal Regulations, Title 21, Parts 800 to 898. In addition, FDA may publish further announcements concerning your device in the Federal Register.

U.S. Food & Drug Administration
10903 New Hampshire Avenue
Silver Spring, MD 20993
www.fda.gov

{1}

K260086 - MinGi Seo

Page 2

Additional information about changes that may require a new premarket notification are provided in the FDA guidance documents entitled "Deciding When to Submit a 510(k) for a Change to an Existing Device" (https://www.fda.gov/media/99812/download) and "Deciding When to Submit a 510(k) for a Software Change to an Existing Device" (https://www.fda.gov/media/99785/download).

Your device is also subject to, among other requirements, the Quality Management System Regulation (QMSR) (21 CFR Part 820), which includes, but is not limited to, ISO 13485 clause 7.3 (Design controls), ISO 13485 clause 8.3 (Nonconforming product), ISO 13485 clause 8.5.2 (Corrective action), and ISO 13485 clause 8.5.3 (Preventative action). Please note that regardless of whether a change requires premarket review, the QMSR requires device manufacturers to review and approve changes to device design and production (ISO 13485 clause 7.3 and ISO 13485 clause 7.5) and document changes and approvals in the Medical Device File (ISO 13485 clause 4.2.3).

Please be advised that FDA's issuance of a substantial equivalence determination does not mean that FDA has made a determination that your device complies with other requirements of the Act or any Federal statutes and regulations administered by other Federal agencies. You must comply with all the Act's requirements, including, but not limited to: registration and listing (21 CFR Part 807); labeling (21 CFR Part 801); medical device reporting (reporting of medical device-related adverse events) (21 CFR Part 803) for devices or postmarketing safety reporting (21 CFR Part 4, Subpart B) for combination products (see https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reporting-combination-products); good manufacturing practice requirements as set forth in the Quality Management System Regulation (QMSR) (21 CFR Part 820) for devices or current good manufacturing practices (21 CFR Part 4, Subpart A) for combination products; and, if applicable, the electronic product radiation control provisions (Sections 531-542 of the Act); 21 CFR Parts 1000-1050.

All medical devices, including Class I and unclassified devices and combination product device constituent parts are required to be in compliance with the final Unique Device Identification System rule ("UDI Rule"). The UDI Rule requires, among other things, that a device bear a unique device identifier (UDI) on its label and package (21 CFR 801.20(a)) unless an exception or alternative applies (21 CFR 801.20(b)) and that the dates on the device label be formatted in accordance with 21 CFR 801.18. The UDI Rule (21 CFR 830.300(a) and 830.320(b)) also requires that certain information be submitted to the Global Unique Device Identification Database (GUDID) (21 CFR Part 830 Subpart E). For additional information on these requirements, please see the UDI System webpage at https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance/unique-device-identification-system-udi-system.

Also, please note the regulation entitled, "Misbranding by reference to premarket notification" (21 CFR 807.97). For questions regarding the reporting of adverse events under the MDR regulation (21 CFR Part 803), please go to https://www.fda.gov/medical-devices/medical-device-safety/medical-device-reporting-mdr-how-report-medical-device-problems.

For comprehensive regulatory information about medical devices and radiation-emitting products, including information about labeling regulations, please see Device Advice (https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-assistance) and CDRH Learn (https://www.fda.gov/training-and-continuing-education/cdrh-learn). Additionally, you may contact the Division of Industry and Consumer Education (DICE) to ask a question about a specific regulatory topic. See the DICE website (https://www.fda.gov/medical-devices/device-advice-comprehensive-regulatory-

{2}

K260086 - MinGi Seo

Page 3

assistance/contact-us-division-industry-and-consumer-education-dice) for more information or contact DICE by email ([e-mail]) or phone [phone] or 301-796-7100).

Sincerely,

Digitally signed by Michael D. O'hara -S

Date: 2026.08.12 13:36:49 -04'00'

For

Yanna Kang, Ph.D.

Assistant Director

Mammography and Ultrasound Team

DHT8C: Division of Radiological

Imaging and Radiation Therapy Devices

OHT8: Office of Radiological Health

Office of Product Evaluation and Quality

Center for Devices and Radiological Health

Enclosure

{3}

| Indications for Use | | |
| --- | --- | --- |
| Please type in the marketing application/submission number, if it is known. This textbox will be left blank for original applications/submissions. | K260086 | ? |
| Please provide the device trade name(s). | | ? |
| CadAI-B Dx | | |
| Please provide your Indications for Use below. | | ? |
| CadAI-B Dx is a software application indicated to assist interpreting physicians in analyzing the breast ultrasound images of patients with soft tissue breast lesions suspicious for breast cancer who are being referred for further diagnostic ultrasound examination. Output of the device includes lesion indicators placed on breast ultrasound images assisting physicians to identify suspicious soft tissue breast lesions from given B-Mode ultrasound images and the risk analysis of malignancy. The risk analysis of malignancy indicates the malignancy score (CadAI-Score) and corresponding BI-RADS categories. In addition, CadAI-B Dx analyzes the size of the lesions and BI-RADS lexicon descriptors: shape, orientation, margin, echo pattern, and posterior features. CadAI-B Dx may also be used as an image viewer of digital ultrasound images. The software includes tools that allow users to adjust, measure and document images, and output into a structured report. Patient management decisions should not be made solely on the basis of analysis by CadAI-B Dx. Limitations: CadAI-B Dx is not to be used on sites of post-surgical excision, or images with Doppler, elastography, or other overlays present in them. | | |
| Please select the types of uses (select one or both, as applicable). | Prescription Use (21 CFR 801 Subpart D)Over-The-Counter Use (21 CFR 801 Subpart C) | ? |

{4}

CADAI-Series Evidence you can get, remember you can do

Beamworks

# 510(k) Summary: K260086

[As Required by 21 CFR 807.92]

## 1. Data Prepared

December 31, 2025

## 2. Submitter's Information

- Name of Manufacturer: BeamWorks Inc.
- Address: B1, 107 Chilgokjungang-daero 136-gil, Buk-gu Daegu, Republic of Korea
- Contact Name: MinGi Seo/Regulatory Affairs
- Telephone No.: [phone]
- Email Address: [e-mail]

## 3. Identification of Proposed Device(s)

| 510(k) Number | K260086 |
| --- | --- |
| Trade/Device Name | CadAI-B Dx |
| Product Name | CadAI-B Dx |
| Regulation Name | Radiological computer assisted detection and diagnosis software |
| Regulation Number | 21 CFR 892.2090 |
| Classification Product Code | QDQ |
| Device Class | Class II |
| 510(k) Review Panel | Radiology |

{5}

CADAI-Series Evidence your company, released we you can be

Beamworks

## 4. Identification of Predicate Device(s)

The identified predicate device within this submission is shown as follows;

| **510(k) Number** | K210670 |
| --- | --- |
| **Trade/Device/Model Name** | BU-CAD |
| **Product Name** | BU-CAD |
| **Regulation Name** | Radiological computer assisted detection and diagnosis software |
| **Regulation Number** | 21 CFR 892.2090 |
| **Classification Product Code** | Classification Product Code: QDQ Subsequent Product Code: LLZ |
| **Device Class** | Class II |
| **510(k) Review Panel** | Radiology |

These predicate devices have not been subject to a design-related recall.

Page 2 / 17

{6}

CADAI-Series Evidence you can get, remember you can do

Beamworks

## 5. Description of the Device

CadAI-B Dx, developed by BeamWorks Inc., is a Software as a Medical Device (SaMD) designed to assist interpreting physicians by providing Computer-Assisted Detection (CADe) and Diagnosis (CADx) capabilities in breast ultrasound examinations. The software employs artificial intelligence and machine learning (AI/ML) algorithms trained on large databases comprising ultrasound images, along with associated metadata such as radiology and pathology reports.

After the clinical acquisition of ultrasound images is complete, the software analyzes the captured B-Mode images, identifies, and highlights suspicious areas of abnormality to assist the users in localization of breast lesions. This functionality directs the user's attention to potential lesions based on objective spatial correspondence criteria, ensuring that AI highlights are aligned with clinically validated detection standards. Simultaneously, the software provides automated measurements of the detected lesions (Longest and Shortest Diameters) and extracts standardized ACR BI-RADS lexicon descriptors (e.g., shape, orientation, margin, echo pattern, and posterior acoustic features) to support diagnostic consistency.

The software generates a quantitative malignancy score (CadAI-Score) on a scale of 0 to 100, which indicates the degree of clinical suspicion based on the lesion characteristics learned by the AI/ML algorithm. This score is calibrated to the American College of Radiology (ACR) BI-RADS system to provide a suggested BI-RADS category (CadAI-BIRADS). To ensure safe and appropriate use, the display of the malignancy score and BI-RADS category is gated via role-based access control; these diagnostic outputs are intended to be accessible only to authorized interpreting physicians qualified to perform diagnostic interpretation.

The analysis results provided by the software can be reviewed, modified, and saved by the physician based on their independent clinical judgment. If a user adjusts the lesion markings or descriptors, the system is designed to automatically recalculate the associated diagnostic values based on the updated inputs. The final analysis results can be converted into DICOM Secondary Capture files and saved to user-designated storage locations, such as a Picture Archiving and Communication System (PACS), to support integrated clinical decision-making.

## 6. Indications for Use

CadAI-B Dx is a software application indicated to assist interpreting physicians in analyzing the breast ultrasound images of patients with soft tissue breast lesions suspicious for breast cancer who are being referred for further diagnostic ultrasound examination.

Output of the device includes lesion indicators placed on breast ultrasound images assisting physicians to identify suspicious soft tissue breast lesions from given B-Mode ultrasound images and the risk analysis of

Page 3 / 17

{7}

CADAI-Series Evidence for Competitive Industries

Beamworks

malignancy. The risk analysis of malignancy indicates the malignancy score (CadAI-Score) and corresponding BI-RADS categories. In addition, CadAI-B Dx analyzes the size of the lesions and BI-RADS lexicon descriptors: shape, orientation, margin, echo pattern, and posterior features.

CadAI-B Dx may also be used as an image viewer of digital ultrasound images. The software includes tools that allow users to adjust, measure and document images, and output into a structured report.

Patient management decisions should not be made solely on the basis of analysis by CadAI-B Dx.

Limitations: CadAI-B Dx is not to be used on sites of post-surgical excision, or images with Doppler, elastography, or other overlays present in them.

## 7. Technological Comparison

Provided below is a table that compares technological characteristics of the CadAI-B Dx and the predicate device.

[Table 3. Comparison of Proposed Device to Predicate Devices]

| | Proposed Device | Predicate Device | Note |
| --- | --- | --- | --- |
| K Number | K260086 | K210670 | - |
| Manufacturer | BeamWorks Inc. | TaiHao Medical Inc. | - |
| Trade Name | CadAI-B Dx | BU-CAD | - |
| Product Name | CadAI-B Dx | BU-CAD | - |
| Product Code | QDQ | QDQ, LLZ | Similar |
| Regulation Number | 21 CFR 892.2090 | 21 CFR 892.2090 | Identical |
| 510(k) Review Panel | Radiology | Radiology | Identical |
| Intended Use | Intended to be used by physician interpreting radiological images, to help them with localizing and characterizing breast abnormalities. | Intended to be used by clinicians interpreting radiological images, to help them with localizing and characterizing breast abnormalities. | Identical |

Page 4 / 17

{8}

CADAI-Series
Evidence your company. Interact with you. Use it.

Beamworks

| | Proposed Device | Predicate Device | Note |
| --- | --- | --- | --- |
| | Intended to be used concurrently with the reading of images and are not intended as a replacement for the review of a physician or their clinical judgement. | Intended to be used concurrently with the reading of images and are not intended as a replacement for the review of a clinician or their clinical judgement. | |
| Indications for Use | CadAI-B Dx is a software application indicated to assist interpreting physicians in analyzing the breast ultrasound images of patients with soft tissue breast lesions suspicious for breast cancer who are being referred for further diagnostic ultrasound examination. Output of the device includes lesion indicators placed on breast ultrasound images assisting physicians to identify suspicious soft tissue breast lesions from given B-Mode ultrasound images and the risk analysis of malignancy. The risk analysis of malignancy indicates the malignancy score (CadAI-Score) and corresponding BI-RADS categories. In addition, CadAI-B Dx analyzes the size of the lesions and BI-RADS lexicon descriptors: shape, orientation, margin, echo pattern, and posterior features. CadAI-B Dx may also be used as | BU-CAD is a software application indicated to assist trained interpreting physicians in analyzing the breast ultrasound images of patients with soft tissue breast lesions suspicious for breast cancer who are being referred for further diagnostic ultrasound examination. Output of the device includes regions of interest (ROIs) and lesion contours placed on breast ultrasound images assisting physicians to identify suspicious soft tissue lesions from up to two orthogonal views of a single lesion, and region-based analysis of lesion malignancy upon the physician's query. The region based analysis indicates the score of lesion characteristics (SLC), and corresponding BI-RADS categories in user-selected ROIs or ROIs automatically identified by the software. In addition, BU-CAD also automatically | Similar |

Page 5 / 17

{9}

CADAI-Series
Excellence your care get, remember you care us

Beamworks

| | Proposed Device | Predicate Device | Note |
| --- | --- | --- | --- |
| | an image viewer of digital ultrasound images. The software includes tools that allow users to adjust, measure and document images, and output into a structured report. Patient management decisions should not be made solely on the basis of analysis by CadAI-B Dx. Limitations: CadAI-B Dx is not to be used on sites of post-surgical excision, or images with Doppler, elastography, or other overlays present in them. | classifies lesion shape, orientation, margin, echo pattern, and posterior features according to BI-RADS descriptors. BU-CAD may also be used as an image viewer of multi-modality digital images, including ultrasound and mammography. The software includes tools that allow users to adjust, measure and document images, and output into a structured report (SR). Patient management decisions should not be made solely on the basis of analysis by BU-CAD. Limitations: BU-CAD is not to be used on sites of post-surgical excision, or images with Doppler, elastography, or other overlays present in them. BU-CAD is not intended for the primary interpretation of digital mammography images. BU-CAD is not intended for use on mobile devices. | |
| Characteristics | CADe and CADx software used to assist in localizing suspicious soft tissue lesions and region | CADe and CADx software used to assist in localizing suspicious soft tissue lesions and region | Identical |

Page 6 / 17

{10}

CADAI-Series Evidence you can get, remember you can be

Beamworks

| | Proposed Device | Predicate Device | Note |
| --- | --- | --- | --- |
| | based analyze of malignancy using ultrasound image data. | based analyze of malignancy using ultrasound image data. | |
| Target Population | Patients with soft tissue breast lesions who are being referred for ultrasound interpreting. | Patients with soft tissue breast lesions who are being referred for ultrasound interpreting. | Identical |
| Anatomical Location | Breast | Breast | Identical |
| Design | Software-only device | Software-only device | Identical |
| Modality Used for Analysis | Breast ultrasound data | Breast ultrasound data | Identical |
| Input | Medical images provided in a DICOM and Image format | Medical images provided in a DICOM format | Similar |
| Output | ROIs (CadAI-Map) and lesion contours placed on soft tissue breast lesions. A region-based malignancy score (CadAI-Score), a BI-RADS category, lesion size, and BI-RADS lexicon descriptors (shape, orientation, margin, echo pattern, posterior features). | ROIs and lesion contours placed on suspicious soft tissue lesion. A region-based score of lesion malignancy, a BI-RADS category, and BI-RADS descriptors. | Similar |
| Physical Characteristics | Software Package Operates on off-the shelf hardware | Software Package Operates on off-the shelf hardware | Identical |
| Comparative Performance Testing (MRMC) | 1. Metric: AUC_LROC2. Cases used: 797● 447 Benign● 350 Malignant3. Readers: 16 | 1. Metric: AUC_LROC2. Cases used: 628● 374 Benign● 254 Malignant3. Readers: 16● 14 Radiologists● 2 Breast Surgeons | Similar |

Page 7 / 17

{11}

CADAI-Series
Evidence you can get, remember you can be

Beamworks

| | Proposed Device | Predicate Device | Note |
| --- | --- | --- | --- |
| | - 5 board-certified radiologists with breast imaging fellowship training. - 3 Breast Surgeons - 5 board-certified radiologists without breast imaging fellowship training. - 6 board-certified physicians from specialties other than radiology fulfills the qualification requirements set by the American College of Radiology (ACR). 4. Results: - AUC improved by 0.107 (0.766 Unaided vs 0.873 Aided) | 4. Results: - AUC improved by 0.037 (0.779 Unaided vs 0.816 Aided) | |
| Performance Testing (Standalone) | 1. Metric: AUC 2. Cases used: 1,285 - 603 Benign - 682 Malignant 3. Source of Cases - South Korea: n=530 (41.25%) - United States: n=643 (50.04%) - Kazakhstan: n=112 (8.72%) 4. Performance (Frequency) - AUC: 0.907 - Sensitivity: 88.12% (601/682) | 1. Metric: AUC 2. Cases used: 1,139 - 642 Benign - 497 Malignant 3. Source of Cases - United States: 531 cases - Europe: 36 cases - Taiwan: 572 cases 4. Performance (Frequency) - AUC: 0.82 - Sensitivity: 88.3% (439/497) - Specificity: 57.9% (372/642) | Similar |

Page 8 / 17

{12}

CADAI-Series Evidence you can get, remember you can do

Beamworks

| | Proposed Device | Predicate Device | Note |
| --- | --- | --- | --- |
| | ● Specificity: 77.11% (465/603) | | |
| Modality Used for Viewing | Breast Ultrasound | Breast Ultrasound and Mammography (FFDM) | Different |

#### - Intended Use

The intended use of CadAI-B Dx is the same as that of the legally marketed predicate device, BU-CAD. Both are intended to be used by physicians interpreting radiological images, to help them with localizing and characterizing breast abnormalities. CadAI-B Dx and the predicate device are both intended to be used concurrently with the reading of images and are not intended as a replacement for the review of a physician or their clinical judgment.

#### ● Indications for Use

Both CadAI-B Dx and BU-CAD are software applications indicated to assist physicians in analyzing breast ultrasound images of patients who are being referred for diagnostic ultrasound examination. Both devices identify regions suspicious for breast cancer and provide computer analytics, including malignancy scores (CadAI-Score for CadAI-B Dx and SLC for BU-CAD) and corresponding BI-RADS categories. Furthermore, both devices automatically classify BI-RADS descriptors such as shape, orientation, margin, echo pattern, and posterior features.

#### ● Intended Use Population and Modality

CadAI-B Dx and BU-CAD share identical intended use populations and modality requirements. Both are intended to be used for assisting interpreting physicians in analyzing patients with soft tissue breast lesions using breast ultrasound data. While BU-CAD may also be used as a viewer for multi-modality images including mammography, both devices are strictly not intended for the primary interpretation of digital mammography images.

#### - Input

Both CadAI-B Dx and the predicate device, BU-CAD, are designed to perform AI-based analysis on a static B-Mode ultrasound image. CadAI-B Dx processes the B-Mode ultrasound images in DICOM and image (RGB) formats.

Page 9 / 17

{13}

CADAI-Series Evidence you can get, remember you can do

Beamworks

# Output

The outputs of CadAI-B Dx and BU-CAD are substantially equivalent. Both devices provide highlighted locations consisting of ROIs and lesion contours placed on suspicious soft tissue lesions. The output for both includes a region-based malignancy score, a BI-RADS category, and BI-RADS descriptors. Additionally, CadAI-B Dx provides lesion size analysis to further assist the physician's diagnostic process.

# Interface

Both CadAI-B Dx and BU-CAD are software-only devices that operate on off-the-shelf hardware. They are intended to be used as image viewers that include tools allowing users to adjust, measure, and document images, subsequently outputting the findings into structured reports or DICOM formats.

# Performance Testing

The clinical validation of CadAI-B Dx followed similar endpoints to those used for BU-CAD. Both devices were evaluated using a Multiple Reader Multiple Case (MRMC) study design with AUC_LROC as the primary metric. The CadAI-B Dx MRMC study evaluated a total of 797 cases (447 benign, 350 malignant) with 16 readers, while the BU-CAD MRMC study evaluated 628 cases with 16 readers. In standalone performance testing, CadAI-B Dx demonstrated robust diagnostic accuracy with an AUC evaluated over 1,285 cases from diverse geographic sources (US, Korea, and Kazakhstan), which is comparable to the 1,139 cases used for BU-CAD's standalone validation. These results demonstrate that CadAI-B Dx performs with substantial equivalence to the predicate device in improving or maintaining reader performance.

# Discussion of the Comparison to Support Substantial Equivalence (SE) Determination

CadAI-B Dx has the same intended use as the legally marketed predicate device, BU-CAD. They are intended to be used by physicians interpreting radiological images, to help them with localizing and characterizing breast abnormalities.

CadAI-B Dx analyzes a static B-Mode ultrasound image, making the analytical input functionally equivalent to the input image used by the predicate. The output of CadAI-B Dx is similar to that of the predicate device by providing ROIs and lesion contours placed on suspicious soft tissue lesions and region-based malignancy scores (CadAI-Score), while providing BI-RADS category and BI-RADS descriptors (Shape, Orientation, Margin, Echo Pattern, and Posterior Features) is also

Page 10 / 17

{14}

CADAI-Series Evidence you can get, remember you can do

Beamworks

consistent with the predicate.

CadAI-B Dx has an identical intended use, compared to the predicate device, that aims to localize and characterize suspicious soft tissue lesions in breast ultrasound image. Artificial intelligence algorithms of the subject device may have different technological characteristics from the predicate device. Therefore, a fully crossed multiple reader multiple case (MRMC) reader study was conducted.

Compared to BU-CAD as the primary predicate, and in consideration of the technological characteristics and test methods used in the clinical validation, CadAI-B Dx does not raise different questions of safety and effectiveness.

## 8. Clinical Performance Data

### 8.1. Summary of the Reader Study

The results of this clinical study demonstrate that CadAI-B Dx provides statistically significant improvements in localization-sensitive clinical effectiveness by assisting physicians in both the accurate classification of malignant lesions and the precise identification of lesion location during breast ultrasound interpretation. Notably, the observed reduction in false-positive interpretations, which addresses a well-recognized limitation of breast ultrasound, together with improved inter-reader agreement supporting diagnostic standardization and meaningful gains in reading efficiency, collectively suggest that CadAI-B Dx offers clinical value beyond improvements in performance metrics alone. These findings support the potential of CadAI-B Dx to function as an effective diagnostic aid that enhances clinical decision-making and improves workflow efficiency in real-world breast ultrasound practice.

#### Dataset Demographic

A total of 797 cases collected from multiple clinical sites were used in the reader study. The source of cases is listed below.

U.S.: 420 cases
● South Korea: 377 cases

The age distribution included in this study was as follows:

● 22–39 years: 92 cases
● 40–49 years: 220 cases

Page 11 / 17

{15}

CADAI-Series
Evidence your company. Interact with you. Get on.

Beamworks

- 50–59 years: 211 cases
- 60–69 years: 154 cases
- ≥70 years: 120 cases

The number of benign and malignant cases included in this study were listed below.

- Benign cases: 447 cases
- Malignant cases: 350 cases

The race distribution included in this study was as follows:

- Asian: 401 cases
- White: 208 cases
- Hispanic: 112 cases
- Black or African American: 71 cases
- American Indian / Alaska Native: 5 cases

The lesion size distribution included in this study was listed below:

- <10 mm: 324 cases
- 10–20 mm: 364 cases
- >20 mm: 109 cases

The malignancy subtypes included in this study were listed below.

- Invasive ductal carcinoma (IDC): 266 cases
- Invasive lobular carcinoma (ILC): 26 cases
- Invasive carcinoma, NOS: 12 cases
- Ductal carcinoma in situ (DCIS): 24 cases
- Other malignant types: 22 cases

The imaging hardware distribution included in this study were listed below:

- GE Healthcare: 391 cases
- Philips: 304 cases
- Canon/Toshiba: 50 cases
- Siemens Healthineers: 45 cases
- Other manufacturers: 7 cases

Page 12 / 17

{16}

CADAI-Series Evidence you can get, remember you can do

Beamworks

The image acquisition years included in this study were listed below:

● 2011–2015: 219 cases
● 2016–2020: 436 cases
● 2021–2023: 142 cases

### Primary Endpoint

Reader-averaged localization and diagnostic performance, measured by AULROC, was higher in the AI-aided reading condition compared with the unaided condition. The reader-averaged AULROC increased from 0.766 (95% CI: 0.725–0.808) in the unaided condition to 0.873 (95% CI: 0.847–0.900) in the AI-aided condition. The difference in reader-averaged AULROC between conditions was 0.107 (95% CI: 0.071–0.142), which was statistically significant (p < 0.001).

Table. Reader-averaged AULROC for Unaided and AI-Aided Sessions

| Metric | Unaided | AI-aided | Aided – Unaided | p-value |
| --- | --- | --- | --- | --- |
| AULROC | 0.766 | 0.873 | 0.106 | < 0.001 |
| 95% CI | 0.725, 0.808 | 0.847, 0.900 | 0.071, 0.142 | |

Note: AULROC = Area under the localization receiver operating characteristic curve

### Diagnostic Accuracy Metrics (Sensitivity, Specificity, PPV, NPV)

Reader-averaged diagnostic accuracy metrics, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), were higher in the AI-aided reading condition compared with the unaided condition.

Table. Reader-averaged Diagnostic Accuracy Metrics

| Metric | Unaided (%) | AI-aided (%) | ΔAided – Unaided (%) | p-value |
| --- | --- | --- | --- | --- |
| Sensitivity | 91.29(89.94, 92.74) | 94.23(92.37, 96.10) | 2.94 | 0.015 |
| Specificity | 36.09(33.67, 38.50) | 55.34(52.21, 58.48) | 19.25 | < 0.001 |
| PPV (unadjusted) | 52.79(49.16, 56.43) | 62.29(58.59, 66.00) | 9.50 | < 0.001 |
| PPV_U.S.(adjusted) | 77.76(76.57, 78.96) | 83.78(82.46, 85.54) | 6.02 (04.46, 07.47) | |
| PPV_Korea(adjusted) | 48.78(47.08,50.53) | 58.45(56.15, 61.69) | 9.67 (07.14, 12.31) | |

Page 13 / 17

{17}

CADAI-Series Evidence you can get, remember you can be

Beamworks

| NPV (unadjusted) | 84.10(81.06, 87.14) | 92.46(89.96, 94.95) | 8.36 | < 0.001 |
| --- | --- | --- | --- | --- |
| NPV_U.S.(adjusted) | 62.85(56.34, 69.90) | 79.67(74.30, 85.45) | 16.82 (10.85, 22.69) | |
| NPV_Korea(adjusted) | 86.13(82.57,89.50) | 93.50(91.39, 95.57) | 7.37 (4.57, 10.32) | |

Note: PPV = positive predictive value; NPV = negative predictive value. PPV and NPV were adjusted using 5-year breast cancer prevalence rates reported by WHO GLOBOCAN 2022 (0.71% for the U.S. and 0.44% for South Korea).

### Subgroup Analysis

Age, race, geographic data source, ultrasound system manufacturer, BI-RADS assessment category, and lesion- and acquisition-related characteristics were evaluated as subgroups using AULROC-based performance metrics. Except for a limited number of subgroups with insufficient sample sizes, AI-aided interpretation with CadAI-B Dx generally demonstrated improved diagnostic performance compared with unaided interpretation. These performance improvements were observed consistently across patient characteristics and technical conditions, supporting the robustness and generalizability of CadAI-B Dx.

## 9. Non-Clinical Performance Data

### 9.1. Summary of the Non-Clinical Bench Testing

The non-clinical bench performance testing provides comprehensive and objective evidence that CadAI-B Dx performs as intended for its labeled indications. The evaluation methodology was designed and conducted in accordance with applicable FDA guidance and relevant recognized standards. The test results demonstrate that the device meets all pre-specified performance acceptance criteria and support reasonable generalizability to the intended use population.

Based on the totality of the non-clinical bench performance evidence presented in this section, CadAI-B Dx is considered substantially equivalent to the predicate device (BU-CAD, K210670) and does not raise new questions of safety or effectiveness when used as a standalone computer-assisted detection and diagnosis tool for breast ultrasound imaging.

### Dataset Demographic

A total of 1,285 cases were used for the standalone performance evaluation. The source of cases is listed below:

● United States: 643 cases
● South Korea: 530 cases
● Kazakhstan: 112 cases

Page 14 / 17

{18}

CADAI-Series Evidence you can get, remember you can do

Beamworks

The age distribution included in this study was as follows:

● 22–39 years: 118 cases
● 40–49 years: 338 cases
● 50–59 years: 372 cases
● 60–69 years: 266 cases
- ≥70 years: 191 cases

The number of benign and malignant cases included in this study were listed below:

- Benign cases: 603 cases
- Malignant cases: 682 cases

The race distribution included in this study was as follows:

- Asian: 556 cases
- White: 412 cases
- Hispanic: 112 cases
- Black or African American: 77 cases
- Other/Not Available: 128 cases

The imaging hardware distribution included in this study were listed below:

● GE Healthcare: 649 cases
● Philips: 426 cases
- Mindray: 112 cases
● Siemens Healthineers: 28 cases
- Others: 69 cases

### Test Dataset Independence

The test dataset (N = 1,285) was strictly sequestered from algorithm development. No data from any patient in the test dataset were used for training, tuning, or internal validation, and no institution that contributed training or tuning data was included in the test dataset. Performance was therefore evaluated on an independent dataset.

### Reference ground truth

A multi-step expert consensus ground truth dataset was established by three independent, MQSA-certified breast imaging radiologists, each with over 10 years of dedicated clinical experience and including active members of the ACR BI-RADS Committee.

Page 15 / 17

{19}

CADAI-Series Evidence you can get, remember you can do

Beamworks

### Primary Performance Analysis

The standalone integrated lesion localization and diagnostic performance, measured by AULROC, met the pre-specified acceptance criterion derived from the predicate device. The observed AULROC was 0.907 (95% CI: 0.890–0.924), with the lower bound of the 95% confidence interval exceeding the performance goal of 0.7626.

Table. Standalone AULROC Performance

| Metric | Value | 95% CI |
| --- | --- | --- |
| AULROC | 0.907 | 0.890, 0.924 |

Note: AULROC: Area under the localization receiver operating characteristic curve

### Sensitivity, Specificity, Positive Predictive Value (PPV), and Negative Predictive Value (NPV)

Standalone diagnostic accuracy metrics were evaluated with a localization penalty to ensure clinical relevance. The system demonstrated a sensitivity of 88.12% (95% CI: 85.74–90.55) and a specificity of 77.11% (95% CI: 73.73–80.40). All diagnostic metrics met or exceeded the predefined acceptance criteria derived from the predicate device.

Table. Sensitivity, Specificity, PPV, NPV with Localization Penalty

| Metric | Value | Frequency | 95% CI |
| --- | --- | --- | --- |
| Sensitivity (%) | 88.12 | 601/682 | 85.74, 90.55 |
| Specificity (%) | 77.11 | 465/603 | 73.73, 80.40 |
| PPV (%) [unadjusted] | 80.35 | 601/748 | 77.23, 83.10 |
| PPV_U.S. (%) | 2.68 | - | 2.32, 3.11 |
| PPV_Korea (%) | 1.67 | - | 1.45, 1.94 |
| NPV (%) [unadjusted] | 86.59 | 465/537 | 83.64, 89.41 |
| NPV_U.S. (%) | 99.90 | - | 99.89, 99.93 |
| NPV_Korea (%) | 99.94 | - | 99.92, 99.95 |

Note: PPV = positive predictive value; NPV = negative predictive value. PPV and NPV were adjusted using 5-year breast cancer prevalence rates reported by WHO GLOBOCAN 2022 (0.71% for the U.S. and 0.44% for South Korea).

### Summary of Subgroup Analysis

Subgroup analyses across demographics (age, race/ethnicity, geographic data source), disease

Page 16 / 17

{20}

CADAI-Series^{}[] Evidence you can get, remember you can do

Beamworks

characteristics, ultrasound systems, acquisition conditions, and BI-RADS assessment category showed consistent standalone discrimination, with AULROC $\geq 0.800$ in all interpretable prespecified subgroups and no meaningful degradation versus overall performance.

## 10. Conclusion

The manufacturer implemented a comprehensive risk management process in accordance with FDA-recognized standards, including ISO 14971, to identify and mitigate risks associated with CadAI-B Dx. Non-clinical bench testing and clinical performance data demonstrated that CadAI-B Dx performs substantially equivalently to the legally marketed predicate device, BU-CAD (K210670), without raising new questions of safety or effectiveness. Accordingly, CadAI-B Dx is determined to be substantially equivalent to the identified predicate device.

Page 17 / 17